Growth Hormone Deficiency due to p.(Gln467Argfs*64) Mutation in the ARID1B Gene in a Girl with Coffin-Siris Syndrome

• Coffin-Siris syndrome (CSS) is an extremely rare genetic multisystemic disorder.

• The utility of whole-exome sequencing (WES) in recognizing causes of intellectual disability is established.

• Despite its phenotypic heterogeneity, growth failure in CSS should be carefully monitored.

• Patients with CSS and growth failure should be evaluated for growth hormone deficiency and treated promptly.

• Use of CSS-specific growth charts are important to better characterize the growth in this population.

Coffin-Siris syndrome (CSS) (MIM #135900) is a multisystem congenital anomaly syndrome. It is inherited in an autosomal dominant manner but most commonly results from a de novo pathogenic variant. Classically, individuals with CSS have been described with hypo- or aplasia of the distal phalanx or nail of the fifth digit (hence the term “fifth digit syndrome”) and specific facial features such as thick eyebrows, long eyelashes, wide nose base, wide mouth with thick and everted lips, and abnormal position or ear shape.

Other physical features include growth restriction, hypertrichosis or hirsutism, sparse scalp hair, dental anomalies as well as developmental and cognitive delay of varying degree. Accompanying congenital abnormalities may include cardiac, genitourinary, gastrointestinal, and central nervous system. Ophthalmological problems, hearing abnormalities, nutrition difficulties as well as growth failure may also be present [Schrier et al., 2013].

The first description of the syndrome was published by Coffin and Siris in 1970. They described 3 unrelated girls with mental retardation, absent nails of the fifth fingers, and hypoplastic distal phalanges [Coffin and Siris, 1970]. To date, less than 200 cases of genetically confirmed CSS have been clinically reported. Exact prevalence and incidence are not known but the disorder is probably under-recognized (https://www.orpha.net/consor/cgibin/).

According to data generated by the Human Disease Genes website series (www.humandiseasegenes.com), speech delay is the most prominent sign in 93.8%, while an estimated 22.5–25% of affected individuals do not develop verbal language skills [van der Sluijs et al., 2019]. Feeding difficulties may be present up to 69.2%, constipation in 29.2%, and general gastro-intestinal problems in 46.2%. One third (29.7%) of the CSS patients will have seizures at least once in their lifetime and 25.8% will be diagnosed with scoliosis. Almost half of the patients (48.2%) will have vision problems, with myopia being the most common deficit (27.2%), while 20.5% will have hearing loss. Accompanying cardiac anomalies (18.9%), renal anomalies (10.6%), and cryptorchidism (48.1%) may also be present. Agenesis of corpus callosum may be seen in up to 41.2% in brain MRI.

Here, we report a case of a 12-year-5-month-old girl with the clinical features of CSS, epilepsy, severe scoliosis, hypothyroidism, and growth hormone deficiency (GHD) caused by a mutation in the ARID1B gene, identified with next-generation sequencing.

A female child was referred at the age of 21 months for neurodevelopmental delay and dysmorphic features. She is the only child of healthy and unrelated parents, born after an uneventful pregnancy with a birth weight of 2.460 kg at the 37th gestation week. She had no perinatal problems. On physical examination, she was a child with good social skills and mild central hypotonia. Height, weight, and head circumference were between the 25th and 50th percentile. She could stand but didn’t make any attempts for walking. She had no expressive speech but could follow commands and had feeding difficulties. She had a wide nose with a flat nasal bridge, wide mouth, and thick eyebrows but eyelashes and hypertrichosis of the trunk were remarkable. Scalp hair was sparse. She also had abnormalities of the fifth (pinky) fingers (Fig. 1a, b). She developed an early onset scoliosis with a left thoracolumbar curve of 20° at the age of 18 months and was wearing a bracing. Clinical picture was compatible with psychomotor retardation, without autistic characteristics. Brain MRI revealed hypoplasia of the corpus callosum. Spinal cord MRI, abdominal ultrasound, cardiological assessment, screening tests for metabolic disorders, and audiological test did not reveal any pathological findings.

During re-examination, at the age of 2 years, she walked independently with physiotherapy intervention and began to use simple words. She completed the process of toilet-training at the age of 4 years. At the age of 5, she still had immature behavior for her age and despite speech therapy, her language development remained delayed (poor vocabulary, short sentences). At that age she had an episode of generalized tonic-clonic seizures. Electroencephalograph (EEG) was abnormal with the presence of epileptiform activity of the right hemisphere (Fig. 2a). She was free of episodes for more than a year when she presented with a similar episode and treatment with valproic acid was then started.

The patient was addressed for short stature evaluation for the first time in the endocrinology department at the age of 9 years and 8 months, showing a remarkable growth retardation. Her height was 124.3 cm (<3th centile) and her weight was 29.65 kg (25th centile). The growth velocity rate was 4.1 cm/year (<2 SDS). She reported being treated for hypothyroidism since the age of 7 years and 4 months and was euthyroid the last few years. The maternal height is 169 cm and the paternal height is 180 cm (both between the 75–90th centile). The target height for the patient was estimated to be 168 cm (± 4.5) (75–90th centile). Left hand’s X-ray confirmed a bone age retardation of 2 years (Fig. 3). Routine blood tests were within normal range. Extensive biochemical and hormonal investigations revealed GHD in 2 stimulation tests. Brain MRI did not reveal pathological findings in the hypothalamic pituitary region. Recombinant human growth hormone (rhGH) treatment was started that resulted in a significant improvement of the growth velocity up to 5.4 cm/year (>90-97th centile). The patient is under treatment until now and presents a nice growth response (Fig. 4).

Until the age of 10 years, she presented scoliosis that was treated conservatively with bracing with a good outcome. After the age of 10 years, there was no compliance due to her neurological problems, and the scoliosis progressed with a right thoracic curve of 47° and a left lumbar curve of 46° (Fig. 5a, b). Very recently, at the age of 12 years, surgical correction of her scoliosis has improved her growth status.

Chromosome analysis from peripheral blood (46,XX) and CGH-array were performed at the age of 2 years and did not reveal abnormalities. Because of the combination of dysmorphology, epilepsy, and intellectual disability, at the age of 10 years, clinical exome sequencing with DNA isolation from the patient was performed. Afterwards, oligonucleotide-based target capture analysis and nucleotide sequencing was performed using the Blueprint Genetics (BpG) Beyond Paediatric Epilepsy Panel Plus and next-generation sequencing (Illumina NextSeq), respectively, examining 283 disease-causing genes. The software Burrows-Wheeler Aligner (BWA-MEM) was used for the data analysis (reference genome UCSC hg19 and reference database Human Gene Mutation Database [HGMD v.2017.1]). Nucleotide change c.1392_1402del, p.(Gln467Argfs*64) was identified in the ARID1B gene at a heterozygotic state associated with the Coffin-Siris Syndrome (Fig. 6). The variation is present in clinical databases as pathogenic (ClinVar: 429976 [Pathogenic**]. dbSNP [151]: rs1131691706 [validated dbSNP entry–Clinical significance: CLIN_pathogenic]). Parental molecular testing was not performed.

Currently, our patient is 12 years and 5 months old. Weight and head circumference are between the 25th and 50th percentile and her height is at the 10th percentile, astonishingly improved since our first documentation before surgery. She has myopia (2.5 dioptries on both eyes) and still has her deciduous teeth (delayed dentition). She attends primary school and has learning disability, but her performance is satisfactory with intervention program. She has very good social behavior, is cooperative, and pleasant. The overall picture of the child is evolutionarily improving.

Even though she is free of seizures for 5 years, her EEG is pathological and she is still on sodium valproate (30 mg/kg/D) (Fig. 2b). For the last 2 years she has been also monitored by the nephrology department because of primary nocturnal enuresis, which is gradually improving without medical treatment.

Before the molecular basis was known, the diagnosis of CSS was based on clinical findings. Formal diagnostic criteria for Coffin-Siris syndrome have not been established; however, the presence or absence of common signs and symptoms in the individual are useful in making a clinical diagnosis. Mutations in a number of different genes, such as ARID1A, ARID1B, ARID2, SMARCA4, SMARCB1, SMARCC2, SMARCE1, and SOX11 have been found to cause CSS. Mutations in DPF2 have also recently been described in individuals with a “Coffin-Siris like” phenotype (https://rarediseases.org/rare-diseases/coffin-siris-syndrome/)(https://omim.org/entry/135900) [Schrier et al., 2018]. The clinical features combined with the findings from molecular testing in our patient are consistent with the diagnosis of ARID1B-related Coffin-Siris syndrome (ARID1B-CSS).

Pathogenic variants in ARID1B are one of the most frequent causes of intellectual disability (ID) as determined by large-scale exome sequencing studies. ARID1B is by far the most frequently mutated gene (51–75%) in CSS [Halgren et al., 2012; Wieczorek et al., 2013; Celen et al., 2017; van der Sluijs et al., 2019]. Large-scale exome sequencing studies invariably find that pathogenic variants in ARID1B are among the most frequently identified causes in unspecified ID cohorts (usually around 1%) [Pranckeniene, 2019; van der Sluijs, 2019]

In a recent study of van der Sluijs et al. [2019], a comparison between individuals with a heterozygous pathogenic variant in ARID1B with an a priori clinical diagnosis of CSS and a group without the a priori clinical diagnosis suggests that apart from CSS-specific dysmorphic features (hypo/aplasia of the fifth digits or nails of the hands/feet, sparse scalp hair, coarse facial features, hypertrichosis) there do not appear to be specific genotype-phenotype correlations between the 2 groups. To date, approximately 100 individuals who do not have the classic CSS phenotype have been identified with a heterozygous pathogenic variant in ARID1B [Santen and Clayton-Smith, 2014; Bramswig et al., 2017; Mannino et al., 2018; van der Sluijs et al., 2019]. In addition, the identification of a pathogenic variant in ARID1B in some members of a large cohort with intellectual disability suggests that the prevalence of pathogenic variants in genes associated with CSS (and possibly of subtle phenotypic features of CSS) may be higher than currently appreciated among those with intellectual disability.

Individuals with pathogenic ARID1B variants are typically at the milder end of the spectrum of CSS and often have normal growth. Consistent with previous reports, our CSS patient exhibited near normal growth parameters at birth [Schrier Vergano et al., 2013]. However, shortly after birth, global growth velocity appeared slower than that of the general population. By the age of 10, the weight and head circumference of the CSS cohort began to approach normal parameters as was the case of our patient. Stature, however, remained shortened (<3th centile) at 10 years of age. Finkielstain et al. [2009] have suggested that thousands of genes work in a coordinated effort across body systems to regulate the postnatal growth process. As genes in the BAF pathway and the SOX family of transcription regulating genes have been shown to play important roles regulating the development of several human tissue types, it is possible that pathogenic variants such as those that cause CSS may interfere with growth in this population [Ho and Crabtree, 2010; Lui et al., 2010; Ronan et al., 2013]. Specifically, mutations in ARID1B have been linked to short stature in both patients with CSS and individuals with non-syndromic ARID1B mutations, implicating this gene as causative for short stature [Santen and Clayton-Smith, 2014; Yu et al., 2015]. Information on the prevalence of comorbid endocrinologic and metabolic abnormalities, however, is limited. Prior to the discovery of the molecular etiology of CSS, several case reports of patients with a clinical diagnosis of CSS cited endocrinologic and metabolic abnormalities that have the potential to influence growth. These case reports included 2 siblings with hypophosphatasia and high alkaline phosphate levels, 1 patient with recurrent hypoglycemia, and 1 patient with GHD that was successfully treated with recombinant GH therapy [Rabe et al., 1991; Imaizumi, 1995; Baban et al., 2008]. In 2017, Celen et al. [2017] reported on 3 individuals (2 with a mutation in ARID1B and 1 with a mutation in SMARCA4) who experienced deficiency in the GHIGF1 (Insulin-like growth factor) axis. Each of these patients reportedly benefited from GH treatment. Later on, in 2019, van der Sluijs et al. [2019] reported detailed information on endocrinological abnormalities in a population of 143 CSS patients. Hypothyroidism was documented in 19%, as in our patient, whereas 1 patient of their cohort had hypothyroidism for 2 years, which spontaneously resolved. 13.7% were diagnosed with GH deficiency, and 6 of those received GH supplementation. They had no information on whether GH deficiency was partial or complete. Likewise, several patients with a growth delay due to GH deficiency were described [Yu at al. 2015], and a similar phenomenon was replicated in an Arid1b heterozygous mouse model [Celen et al., 2017]. Considering 30% of patients had short stature (<–2 SDS) [Pedicelli et al., 2009], GH deficiency could be an underrecognized feature of ARID1B patients. Although short stature and delayed bone age have been frequently reported in the literature (almost 50% of patients), no hormonal abnormalities have been demonstrated consistently.

Notably, individuals with CSS often exhibit delayed bone age of about 2–3 years behind the chronological age [Schrier Vergano et al., 2013], as was the case with our patient. Impaired growth has also been observed in Arid1b heterozygous mice [Celen et al., 2017]. These growth differences occurred in the context of equivalent nutritional intake compared to wild-type mice. Upon further investigation, the mice were found to have insufficient plasma levels of IGF1 with poor GH and growth hormone-releasing hormone (GHRH) compensation. Treatment with recombinant mouse GH increased weight, length, and muscle weakness, further demonstrating the potential for the use of GH to treat some physical differences associated with ARID1B mutations. Further research will be needed to determine the effectiveness of GH supplementation for improving growth in the broader CSS population. Very recently a study was conducted [McCague et al., 2020] to construct CSS-specific growth charts to better characterize the growth in this population. Anthropometric data were collected from 99 individuals enrolled in the CSS/BAF pathway international registry via a retrospective chart review. Normative growth curves are beneficial in gauging whether or not a nutritional intervention is necessary and for future research into the need of GHD evaluation. Growth curves are also valuable for surveillance, epidemiological research, and clinical use. We recommend utilizing growth curves in the clinical setting to best determine nutritional needs and appropriate interventions to optimize the growth potential of patients with CSS. In addition, these growth curves are valuable for future research evaluating the impact of GH supplementation and other novel interventions to support growth and development in this population. As for the individuals in the cohort of McCague et al. [2020] known to have received GH therapy, only one individual was diagnosed with isolated pediatric GHD due to the ectopic pituitary gland without other pituitary dysfunction and has seen improved height with GH treatment. Additional information was unavailable for the other 2 individuals.

Moderately feeding problems are noted in two thirds and Hypoplasia of the corpus callosum in one third. Approximately one third of individuals will experience seizures, predominately of the tonic-clonic type, and appear to respond well to standard antiepileptic drugs (AEDs) as our patient did. No one AED has been demonstrated effective specifically for this disorder. Additional individuals may have abnormal EEGs (up to 6%) without apparent clinical seizure activity. The age of onset of seizures ranges from birth to mid-teenage years, with a median of 4 years [van der Sluijs et al., 2019]. Scoliosis also occurs in 30–40% during infancy. Delayed dentition has also been described, as in our patient. Myopia is the most reported vision problem (27.5%). Facial gestalt is consistent with CSS, albeit at times milder, with hypertelorism and anteverted nares more commonly noted. Distal digital hypoplasia is usually limited to the fifth digit [Van der Sluijs et al., 2019]

CSS appears to follow an autosomal dominant pattern of inheritance. With the exception of 2 families in which a parent and child had features consistent with ARID1B-related disorder, all documented pathogenic ARID1B mutations to date have arisen in a sporadic, de novo manner [van der Sluijs, 2019]. Prenatal testing is technically feasible, but the likelihood of recurrence in families who have had an affected child is considered low based upon the current knowledge.

Herein, we report a 12-year-5-month-old girl  with psychomotor retardation, specific facial features, epilepsy, growth failure due to GH deficiency, severe scoliosis, and delayed dentition compatible with CSS, caused by ARID1B gene mutation, identified with next-generation sequencing.

As described in the literature, despite its phenotypic heterogeneity, key features of CSS have become clearer and along with molecular diagnosis a further global approach to improve the care of these individuals is enabled. Appropriate therapies for this population are needed to optimize growth and intellectual potentials.

We thank the patient’s family for kindly participating in this article, and all the physicians that collaborated in diagnosis and clinical follow-up of the patient.

This case report was produced in accordance with institutional policies; the patient’s parents gave written, informed consent for publication of this case report and any accompanying images.

The paper is exempt from ethical committee approval, since this is a retrospective presentation of the case report.

Ethical approval was not required for this study in accordance with local/national guidelines.

The authors have no conflicts of interest to declare.

No funding sources to report.

All authors read and approved the final version of the manuscript. S.Mo. was in charge of overall patient management, preparation, and finalization of the manuscript providing important intellectual inputs. S.L.-E. was in charge of the endocrinology follow-up and prepared and finalized the endocrinology part in the manuscript. S.Ma. critically read the manuscript. I.P. and E.M. performed and analyzed the next-generation sequencing data and the analysis of the identified variants. N.S. is the treating orthopedic doctor. A.Ko. and A.Ka. searched the literature. E.K.-P. and K.V. supervised patient work-up and management.

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.